Abstract
The aim is to develop new traffic detection system for adaptive control of junctions and traffic surveys, which will combine the advantages of radar and optical technology using artificial intelligence methods both in image processing and aggregation of individual measurements. The detector computation unit will be based on new AI processors for embedded systems that allow deep neural networks and probabilistic object tracking to run. The resulting sensor will have a defined REST interface for integration into higher units. The sensor will also be equipped with a 4G / 5G modem for data connectivity, GPS locator and memory storage for recording and storing up to 3 weeks of data.